Physico-chemical functioning and development of phytoplankton in Karaoun reservoir (Lebanon) : application of a hydrodynamic-ecological model

HAL Id: tel-01127361

https://pastel.archives-ouvertes.fr/tel-01127361

Submitted on 7 Mar 2015

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Physico-chemical functioning and development of

phytoplankton in Karaoun reservoir (Lebanon) :

application of a hydrodynamic-ecological model

Ali Fadel

To cite this version:

Ali Fadel. Physico-chemical functioning and development of phytoplankton in Karaoun reservoir (Lebanon) : application of a hydrodynamic-ecological model. Hydrology. Université Paris-Est, 2014. English. �NNT : 2014PEST1064�. �tel-01127361�

A Dissertation presented to obtain Doctoral degree from

Université Paris-Est

Speciality: Sciences and Techniques of Environment

by

Ali FADEL

Doctoral School: Sciences, Engineering and Environment

Physico-chemical functioning and development of phytoplankton in

Karaoun Reservoir (Lebanon). Application of a

hydrodynamic-ecological model.

This thesis was conducted between École Nationale des Ponts et Chaussées (France) and Commission Libanaise de l'Energie Atomique (Lebanon) and defended on 22 September

2014 in front of the thesis committee composed of: Myriam Bormans Reviewer

Catherine Quiblier Reviewer

Sarah Dorner Examiner

Julien Némery Examiner

Brigitte Vinçon-Leite Thesis director Kamal Slim Thesis director

Bruno Lemaire Thesis co-director

Ali Atoui Thesis co-director

Bruno Tassin Invited member

Nabil Amacha Invited member

Thèse présentée pour obtenir le grade de

docteur de l’ :

Université Paris-Est

Spécialité : Sciences et Techniques de l’Environnement

par

Ali FADEL

Ecole Doctorale : Sciences, Ingénierie et Environnement

Fonctionnement physico-chimique et développement du

phytoplancton dans le réservoir de Karaoun (Liban). Application

d’un modèle couplé hydrodynamique-écologique.

Thèse soutenue le 22 septembre 2014 devant le jury composé de :

Myriam Bormans Directeur de recherche, Université de Rennes 1 (Rapporteur)

Catherine Quiblier Maître de conférences HDR, Université Paris Diderot (Rapporteur)

Sarah Dorner Professeur, École Polytechnique Montréal (Examinateur)

Julien Némery Enseignant-chercheur, Université de Grenoble (Examinateur)

Brigitte Vinçon-Leite Charge de recherche HDR - École Nationale des Ponts et Chaussées (Directrice de thèse)

Kamal Slim Professeur - Commission Libanaise de l'Energie Atomique (Directeur de thèse)

Ali Atoui Directeur de recherche, Commission Libanaise de l'Energie Atomique (CoDirecteur de thèse)

Bruno Lemaire Enseignant-chercheur - École Nationale des Ponts et Chaussées / AgroParisTech (CoDirecteur de thèse)

Bruno Tassin Directeur de recherche - École Nationale des Ponts et Chaussées (Membre invité)

This thesis work was achieved at:

Laboratoire Eau

Environnement et Systèmes

Urbains (LEESU)

&

Lebanese Atomic Energy

Commission (LAEC)

thanks to financial support by:

National Council for Scientific

Research (CNRS), Lebanon

&

École Nationale des Ponts et

Chaussées (ENPC)

ACKNOWLEDGEMENTS

I would like to express my gratitude to all those who helped me during my PhD study. Many people have made invaluable contributions, both directly and indirectly to my research.

First, I would like to thank my advisors, Bruno LEMAIRE, Brigitte VINÇON-LEITE, Kamal SLIM, Ali ATOUI and Bruno TASSIN. I owe you so much. You’ve been my friends, my mentors, my confidants, my colleagues, and moral supporters. You have given so much of yourself to help me succeed.

I am so grateful to Mouin HAMZE, the secretary general of the Lebanese National Council for Scientific Research and Bilal NSOULI, the director of the Lebanese Atomic energy Commission. This thesis included a lot of field and laboratory research. Thank you very much! Without your trust, support and the logistic and financial facilities you provided, I would not have achieved this work.

I would like to express my appreciation to my advisory committee: Ghassan CHEBBO and Cécile BERNARD provided me with advice and comments on my research.

I wish to thank Myriam BORMANS and Catherine QUIBLIER have honored me to kindly participate as reporters of this thesis and Julien Némery and Sarah Dorner for their role as examiners.

I am grateful to Nabil AMACHA and Ghassan EL-ZEIN for their useful discussions about the functioning of Karaoun Reservoir, technical help on field and hydrological and meteorological data supply.

I would also like to express my thanks to Mohamed SAAD and Philippe DUBOIS for their explanation and help both in laboratory and field work during my stay in France.

I would like to thank Maher Matar, the internship student who helped me in 2013 spring-summer field campaigns and laboratory analysis.

I also owe my sincere gratitude to my friends and colleagues who gave me their care and help during my stay in France. They are Mohamad RAMMAL, Ali HANNOUCHE, Guy YOUNES, Yossef NOHRA, Neng, Talita SILVA, Behzad, George FITTON, Viet, Emna, Jérémie, Yacine, Damien, Abdellah, Bachar, Mohammad AL-HAJ, Odissa, Hamouda and Mahmoud. Besides, I wish to thank all the other members of LEESU: Annick, Catherine, Ioulia, Daniel, Marie-Christine, Patrick Elias and the laboratory director Régis Moilleron.

My deepest thanks would go to my beloved parents (Hassan and Salwa), my sisters (Mirna, Mariam and Maya), my brother Mohammad and my sweetheart Rosa. Without your unending support and love, I never would have made it through this process or any of the tough times in my life. Thank you.

This thesis has been supported and funded by Ecole des Ponts ParisTech, as well as by the Lebanese National Council for Scientific Research, the French Ministry for Higher Education

and Research, the French Ministry for Foreign Affairs through the CEDRE program (project 10 EF 38/L9).

Abstract

Forty percent of world reservoirs suffer from eutrophication which increases phytoplankton biomass in reservoirs and impairs water uses. Understanding the mechanisms and processes that control cyanobacterial blooms is of great concern. Ecosystem models enable us to simulate, analyze and understand ecological processes in lakes and reservoirs. Except for Lake Kinneret in Israel, the phytoplankton community and ecological model application are poorly documented in the Middle East. Karaoun Reservoir, also known as Qaroun, Qaraoun or Qarun, the largest water body in Lebanon, was built for irrigation and hydropower production. There is a great interest in the water quality of this reservoir as it will be used to supply the capital Beirut with drinking water.

The main objective of this Ph.D. work is to understand the dynamics of phytoplankton in Karaoun Reservoir. This main objective branches into three sub-objectives: 1) to establish the seasonal phytoplankton succession 2) to understand the cyanobacterial dynamics 3) identify the driving factors of the cyanobacterial blooms.

To achieve these objectives, we conducted sampling campaigns and laboratory measurements were conducted semi-monthly between May 2012 and August 2013 to assess the trophic state and the phytoplankton biodiversity and seasonal dynamics of its phytoplankton community in response to changes in environmental conditions. These field campaign measurements were then used to calibrate (summer and autumn 2012) and validate (spring and summer 2013) a one-dimensional hydrodynamic-ecological model on Karaoun Reservoir. Our results show that:

• Karaoun Reservoir which strongly stratified between May and August was found eutrophic with a low biodiversity, only 30 phytoplankton species in 2012-2013. Comparing its trophic status and biodiversity to other Mediterranean freshwater bodies showed that it matched more with El Gergal Reservoir in Spain than with natural lakes like Lakes Kinneret in Israel and Trichonis in Greece that were oligotrophic with a high biodiversity.

• Thermal stratification established in spring reduced the growth of diatoms and resulted in their replacement by mobile green algae species during high nutrients availability and water temperatures lower than 22 °C. Water temperature higher than 25 °C favours cyanobacterium Microcystis aeruginosa that displaces Aphanizomenon ovalisporum in summer. Dinoflagellate Ceratium hirundinella dominated in mixed conditions, at low light intensity in late autumn at 19 °C.

• Unlike the high temperatures, above 26 °C, which are associated with blooms of Aphanizomenon ovalisporum in Lakes Kinneret (Israel), Lisimachia and Trichonis (Greece) and in Arcos Reservoir (Spain), Aphanizomenon ovalisporum bloomed in Karaoun Reservoir in October 2012 at a relatively low subsurface temperature of 22°C while the reservoir was weakly stratified. This suggests that the risk of Aphanizomenon ovalisporum blooms in European lakes is high.

• Cylindrospermopsin, a fatal toxin, was detected in almost all samples even when Aphanizomenon ovalisporum was not detected. It reached a concentration of 1.7 µg/L, higher than the drinking water guideline value of 1 µg/L of the World Health Organization. The toxin vertical profiles suggest its possible degradation or

sedimentation resulting in its disappearance from water column. Aphanizomenon ovalisporum biovolumes and cylindrospermopsin concentration were not correlated (n = 31, r2 = - 0.05).

• A simple configuration of the one-dimensional hydrodynamic-ecological model DYRESM-CAEDYM successfully simulated the growth and succession of the cyanobacteria Aphanizomenon ovalisporum and Microcystis aeruginosa. The model showed a good performance in simulating the water level (RMSE < 1 m, annual variation of 25 m), water temperature profiles (RMSE < 1.1 °C, range 13-28 °C) and cyanobacteria biomass (RMSE < 57 µg L-1 equivalent chlorophyll a, range 0-206 µg L-1). Implementing the model to better understand the succession between cyanobacteria blooms showed that higher maximum production of Microcystis aeruginosa during favourable temperature and light conditions allowed it to outgrow Aphanizomenon ovalisporum.

On the local scale, this thesis provides important background data for the Lebanese water management authorities who aim to use this reservoir for drinking water production. It also increases the understanding of processes and mechanisms that control cyanobacterial blooms. The application of simple model configurations with few major processes can be transposed on other eutrophic lakes and reservoirs to describe the competition between dominant phytoplankton species, contribute to early warning systems or be used to predict the impact of climate change and management scenarios.

Keywords: Phytoplankton dynamics, Harmful algal blooms, Cyanobacteria, Ecological

Résumé

40 % des retenues dans le monde souffrent d'eutrophisation. L’augmentation de biomasse de phytoplancton et les proliférations de cyanobactéries dans les réservoirs perturbe leurs usages. Comprendre les mécanismes qui contrôlent la prolifération des cyanobactéries est de grande importance. Les modèles d'écosystèmes lacustres nous permettent de simuler, d'analyser et de comprendre les processus écologiques dans les lacs et les réservoirs. La dynamique phytoplanctonique dans les réservoirs du Moyen-Orient est peu documentée jusqu’à présent. Très peu d'applications de modèles d’écosystèmes lacustres y ont été réalisées. Le réservoir de Karaoun, le plus grand au Liban, a été construit pour l'irrigation et la production hydroélectrique. La préservation de la qualité de l’eau de ce réservoir est d’importance majeure car un projet prévoit de l’utiliser pour l’alimentation en eau potable de la capitale Beyrouth à l’horizon 2025.

Les objectifs de la thèse sont de concevoir et réaliser des campagnes de terrain pour suivre et comprendre la dynamique du phytoplancton et des cyanobactéries dans le lac de barrage de Karaoun, puis de modéliser le fonctionnement physique et biogéochimique de cette retenue.

Des campagnes d'échantillonnage ont été effectuées deux fois par mois entre mai 2012 et Août 2013 pour évaluer l'état trophique, la diversité phytoplanctonique et la dynamique du phytoplancton en réponse aux changements des conditions environnementales. Ces mesures ont été ensuite utilisées pour calibrer (été et automne 2012) et valider (printemps et été 2013) un modèle hydrodynamique-écologique unidimensionnel sur du réservoir. Nos résultats ont montré que:

La retenue de Karaoun, fortement stratifiée thermiquement entre mai et août, est eutrophe, et présente une faible diversité phytoplanctonique. Seulement 30 espèces de phytoplancton ont été recensées en 2012-2013.

La stratification thermique qui apparaît au printemps réduit la croissance des diatomées et entraîne leur remplacement par des chlorophycées. Les concentrations en nutriments sont alors élevées et la température de l'eau est inférieure à 22 °C. Les cyanobactéries dominent en été : Aphanizomenon ovalisporum lorsque la température de surface de l'eau est inférieure à 25 °C, Microcystis aeruginosa lorsqu'elle est supérieure à 25°C. Le dinoflagellé Ceratium hirundinella constitue l’espèce dominante en fin d’automne lorsque la colonne d'eau est mélangée, l’intensité lumineuse est faible et la température de l’eau d’environ 19 °C.

Contrairement aux températures de surface élevées, supérieures à 26 °C, auxquelles prolifère A. ovalisporum dans les lacs de Tibériade (Israël), Lisimachia et Trichonis (Grèce) et la retenue d'Arcos (Espagne), une prolifération d’A. ovalisporum survient en octobre 2012 dans la retenue de Karaoun, à une température de l’eau de 22 °C et alors que la stratification thermique est faible.

La cylindrospermopsine (CYN), une toxine qui a provoqué la mort de bétail par ingestion, a été détectée dans la retenue de Karaoun, même en l’absence d’A. ovalisporum, seule espèce qui la produit identifiée dans la retenue. Les biovolumes d’A. ovalisporum et la concentration de CYN ne sont pas corrélés (n = 31, r2 = - 0,05). La CYN atteint une concentration de 1,7 µg/L, supérieure à la valeur guide pour l'eau potable de 1 µg/L (Organisation Mondiale de la Santé). Les profils verticaux de la

toxine suggèrent que sa disparition progressive de la colonne d'eau est due à sa dégradation ou à sa sédimentation.

Une configuration simple du modèle DYRESM-CAEDYM a permis de simuler avec succès la croissance et la succession des cyanobactéries A. ovalisporum et M. aeruginosa. Le modèle réalise de bonnes performances pour la simulation du niveau de l'eau du réservoir (RMSE <1 m pour une variation annuelle de 25 m), des profils de température de la colonne d'eau (RMSE <1 °C pour des variations annuelles comprises entre 13 et 28 °C) et de la biomasse des cyanobactéries (RMSE <48 µ g/L équivalent chlorophylle-a, concentration entre 0 et 206 µg/L).

A l'échelle locale, cette thèse est importante pour les autorités de gestion des eaux libanaises qui visent à utiliser ce réservoir pour la production d'eau potable. Elle a également permis de mieux comprendre les processus qui contrôlent la prolifération de cyanobactéries. L'application de configurations de modèles simples limités aux processus principaux pourrait être transposée sur d'autres réservoirs eutrophes pour décrire la compétition entre les espèces phytoplanctoniques dominantes, s’insérer dans des systèmes d'alerte ou prédire l'impact du changement climatique et tester des scénarios de gestion.

Mots Cles : Dynamique de phytoplancton, Proliférations d'algues toxiques, Cyanobactéries,

TABLE OF CONTENTS

ACKNOWLEDGEMENTS... 7

Abstract... 9

Résumé ... 11

TABLE OF CONTENTS... 13

SYMBOLS AND ABBREVIATIONS ... 17

General Introduction ... 21

Problem statement ... 21

Thesis objectives and methodology... 23

Structure of the thesis ... 23

CHAPTER 1 LITERATURE REVIEW... 25

1.1 Reservoirs and their ecosystems ... 25

1.1.1 Actual and future development of reservoirs ... 25

1.1.2 Differences between lakes and reservoirs ... 27

1.1.3 Physical functioning of lakes and reservoirs... 28

1.1.3.1 Development of thermal stratification ... 28

1.1.3.2 Stratification and mixing patterns in lakes and reservoirs ... 30

1.1.3.3 Impact of stable stratification on water quality and biodiversity... 31

1.2 Cyanobacteria in freshwater bodies... 32

1.2.1 Ecological and health impacts of toxic cyanobacterial blooms... 33

1.2.1.1 Impacts on ecosystems ... 33

1.2.1.2 Toxin production ... 35

1.2.1.2.1 Microcystins ... 37

1.2.1.2.2 Cylindrospermopsin... 38

1.2.2 Main functional traits and key controlling factors of cyanobacterial blooms ... 39

1.2.2.1 Temperature... 39

1.2.2.2 Light ... 40

1.2.2.3 Nutrients ... 40

1.2.2.4 Vertical migration... 41

1.2.2.5 Grazing ... 42

1.2.2.6 Wind mixing and flushing ... 42

1.3 Water quality models and phytoplankton dynamics in reservoirs... 43

1.3.1 Lake ecosystem models and modelling procedure... 43

1.3.2 Overview of the most commonly applied hydrodynamic-ecological models ... 44

1.3.2.1 CAEDYM... 44 1.3.2.2 DELFT3D-ECOLOGY ... 45 1.3.2.3 CE-QUAL-W2 ... 47 1.3.2.4 PROTECH... 47 1.3.2.5 PCLAKE ... 47 1.3.2.6 IPH-TRIM3D-PCLake ... 48 1.3.2.7 MyLake ... 48 1.3.2.8 SALMO ... 48 1.3.2.9 GLM-AED... 49 1.3.2.10 MELODIA... 49

CHAPTER 2 STUDY SITE AND METHODOLOGY ... 51

2.1 Study site ... 51

2.1.1 Geology and hydrology of Karaoun Reservoir ... 52

2.1.1.1 Reservoir geology... 52

2.1.1.2 Reservoir hydrology ... 54

2.1.1.2.1 Reservoir inflows... 54

2.1.1.2.2 Reservoir outflows and losses... 54

2.1.2 Current and anticipated uses of Karaoun Reservoir ... 56

2.1.2.1 Hydropower production... 56

2.1.2.2 Future water supply to Beirut ... 57

2.1.2.3 Irrigation through Canal 900 ... 57

2.1.2.4 Future Canal 800 ... 57

2.1.2.5 Professional fishing ... 58

2.2 Design of a monitoring program ... 58

2.2.1 Field measurements ... 58

2.2.1.1 Water sampling sites... 58

2.2.1.2 Water sampling method... 59

2.2.1.3 Transparency ... 60

2.2.1.4 Phycocyanin profile measurements ... 60

2.2.1.5 Water temperature, pH and conductivity measurements ... 60

2.2.1.6 Dissolved oxygen ... 61

2.2.2 Laboratory analyses ... 61

2.2.2.1 Phytoplankton microscopic identification and counting... 61

2.2.2.2 Chlorophyll-a quantification... 62

2.2.2.3 Nutrient analysis... 62

2.2.2.4 Cylindrospermopsin analysis... 63

2.2.3 Measurements used to validate the model... 63

2.3 Model description ... 63 2.3.1 DYRESM description ... 64 2.3.2 CAEDYM description... 65 2.3.2.1 Growth rate... 66 2.3.2.2 Temperature... 66 2.3.2.3 Light ... 67

2.3.2.4 Cyanobacteria vertical migration... 68

2.3.2.5 Respiration, Mortality & Excretion ... 69

2.3.3 DYRESM-CAEDYM input data... 70

2.4 Evaluation methods ... 71

2.4.1 Phytoplankton biodiversity ... 72

2.4.2 Trophic state... 72

2.4.3 DYRESM-CAEDYM model performance ... 73

CHAPTER 3 EVALUATION OF TROPHIC STATE, BIODIVERSITY AND ENVIRONMENTAL FACTORS ASSOCIATED WITH PHYTOPLANKTON SUCCESSION IN KARAOUN RESERVOIR ... 75

3.1 Introduction ... 75

3.2 Trophic state and algal succession in Karaoun Reservoir before 2012... 76

3.2.1 Nutrient concentrations and trophic state... 76

3.2.2 Algal succession and biodiversity ... 78

TABLE OF CONTENTS 3.3.1 Hydrological conditions ... 82 3.3.2 Physico-chemical parameters... 83 3.3.2.1 Transparency ... 83 3.3.2.2 Dissolved oxygen ... 84 3.3.2.3 Specific conductivity ... 84

3.3.2.4 Water temperature and thermal stratification ... 85

3.3.2.5 Nitrate and ammonium ... 87

3.3.2.6 Total phosphorus and orthophosphate ... 89

3.3.3 Chlorophyll-a and phycocyanin fluorescence ... 91

3.3.4 Phytoplankton composition and biovolumes ... 93

3.3.5 Phytoplankton groups seasonal succession ... 96

3.3.6 Zooplankton community ... 102

3.3.7 Trophic level and diversity index... 102

3.4 Environmental drivers of the succession of phytoplankton groups in Karaoun Reservoir ... 105

3.4.1 Settling of diatoms after establishment of thermal stratification in early spring... 105

3.4.2 Disappearance of green algae after nutrient limitation and temperature elevation in late spring... 105

3.4.3 Cyanobacteria dominance at high temperature and low nutrient concentrations between late spring and early autumn ... 106

3.4.4 Dominance of dinoflagellate at low irradiance and water temperature in autumn ... 107

3.5 Comparison with other Mediterranean lakes and reservoirs ... 107

3.5.1 Morphological and hydrological characteristics ... 108

3.5.2 Eutrophication level and integrated water management ... 109

3.5.3 Phytoplankton diversity ... 110

3.5.4 Toxic cyanobacterial succession ... 111

3.6 Conclusion ... 112

CHAPTER 4 COMPETITION BETWEEN TWO CYANOBACTERIAL SPECIES, APHANIZOMENON OVALISPORUM AND MICROCYSTIS AERUGINOSA IN KARAOUN RESERVOIR, LEBANON ... 115

4.1 Introduction ... 115

4.2 Results... 116

4.2.1 Physico-chemical conditions... 116

4.2.2 Replacement of Aphanizomenon ovalisporum by Microcystis aeruginosa at high temperature... 117

4.2.3 Cylindrospermopsin detection ... 121

4.2.4 Comparison between A. ovalisporum and CYN distribution in the water column ... 122

4.2.5 Absence of correlation between Cylindrospermopsin concentration and A. ovalisporum biovolumes 123 4.3 Discussion ... 123

4.3.1 Aphanizomenon ovalisporum blooms in Karaoun Reservoir ... 123

4.3.2 Competition between Microcystis aeruginosa and Aphanizomenon ovalisporum... 126

4.3.3 Relation between cylindrospermopsin concentrations and A. ovalisporum ... 127

4.3.4 Disappearance of CYN from water column by degradation or sedimentation... 127

4.4 Conclusion ... 128

CHAPTER 5 MODELLING THE SEASONAL COMPETITION BETWEEN TOXIC CYANOBACTERIA MICROCYSTIS AERUGINOSA AND APHANIZOMENON OVALISPORUM... 131

5.1 Introduction ... 131

5.2 Description of input data to DYRESM-CAEDYM ... 132

5.3 DYRESM-CAEDYM configuration... 133

5.4 Thermal model calibration and verification... 138

5.5 Biological model calibration and validation ... 143

5.6 Succession of Aphanizomenon ovalisporum and Microcystis aeruginosa according to DYRESM-CAEDYM... 146 5.7 Model performance ... 148 5.8 Model limitations ... 149 5.9 Conclusion ... 152 General conclusion ... 155 References ... 161 List of figures... 189 List of tables... 193 APPENDICES ... 195

Appendix A: Field measurement and laboratory analysis protocols... 197

I. Chlorophyll-a analysis ... 197

II. Nutrient analysis ... 199

III. Phytoplankton counting ... 211

IV. Cylindrospermopsin analysis by ELISA ... 214

V. Phycocyanin quantification by Trios microflu-blue ... 217

VI. Starmon temperature sensors, measurement protocol ... 221

Appendix B: Additional figures & tables ... 223

SYMBOLS AND ABBREVIATIONS

AEMON: Aquatic Ecosystem Modelling Network ANR: Agence Nationale de la Recherche

CAEDYM: Computational Aquatic Ecosystem DYnamic Model Chl-a: Chlorophyll-a

CTSI: Carlson’s Trophic State Index CWR: Centre for Water Research CYN: Cylindrospermopsin DO: Dissolved Oxygen

DYCD: DYRESM-CAEDYM

DYRESM: DYnamic REservoir Simulation Model EDF: Electricité de France

ELISA: Enzyme-Linked ImmunoSorbent Assay ET: Evapotranspiration

G: Giga (109)

GLEON: Global Lake Ecological Observatory Network GLM: General Lake Model

GranD: Global Reservoir and Dam Database h: Hour

ha: Hectare

HABs: Harmful Algal Blooms

ICOLD: International commission on Large Dams k: Kilo (103)

L: Litre

LAEC: Lebanese Atomic Energy Commision

LHCs: Light Harvesting Chlorophyll-protein complexes M: Mega (106)

m: Metre

MAPE: Mean Absolute Percentage Error MC-LR: Microcystine Leucine-Arginine

µg L-1: Micrograms per Litre

µS cm-1: Micro-siemens per Centimeter

mins: Minutes

mg m-3: Milligrams per Cubic meter

mg L-1: Milligrams per liter

mm: Millimeter n: Number of Samples NO3: Nitrate

SYMBOLS AND ABBREVIATIONS

OD: Optical Density PC: Phycocyanine

PROTECH: Phytoplankton RespOnses To Environmental Change R2:Coefficient of Determination

RMSE: Root Mean Square Error s: Siemens (unit of conductance) TP: Total Phosphorus

T: Water Temperature UV/VIS: Ultraviolet–visible WCD: World Comission on Dams WQI: Water Quality Index

General Introduction

Problem statement

Over 45,000 large dams1 had been constructed in more than 140 countries by the end of the 20th century (WCD (World Comission on Dams), 2000). An ongoing increase in the construction of reservoirs is expected in the future (Seitzinger et al., 2010). These artificial water bodies meet human needs for drinking water supply, agricultural irrigation, power generation, industrial and cooling water supply, commercial fishing and recreation (Jørgensen et al., 2005b).

But fertilizers and untreated sewage in their catchments often increase nutrient concentrations in these ecosystems and cause their eutrophication (Smith and Schindler, 2009). Eutrophication threatens freshwater bodies as it promotes the development and the persistence of harmful algal blooms. Many lakes and reservoirs throughout the world suffer from toxic cyanobacterial blooms (Bormans et al., 2004; Paerl and Paul, 2012). For example, in France, a research project at the national scale showed that all lentic ecosystems can be affected by cyanobacteria blooms (Sarazin et al., 2002). These harmful photosynthetic species reduce ecosystem biodiversity and can produce toxins (neurotoxins, hepatotoxins, cytotoxins, dermotoxins and endotoxins)

.

Some of these toxins cause skin irritation upon contact, illness, intoxication and death to livestock, pets, and wildlife that ingest water contaminated with toxic cyanobacterial cells or toxins released from decaying cyanobacterial cells (Araoz et al., 2010; Briand et al., 2003; Lance et al., 2010; Sotton et al., 2014). In addition, other nuisances are attributable to bloom-forming cyanobacteria. They include: 1) a decrease in water transparency; 2) a reduction in the dissolved oxygen concentration in the hypolimnion; 3) bad smell and scum production; and 4) several negative economic impacts such as preventing the recreational use of the water bodies, clogging irrigation pumps and disturbing hydropower equipment (Smith, 2003).

The regular monitoring of the environmental status of aquatic ecosystems is essential, it is achieved by studying their physico-chemical and biodiversity conditions (Piha and

General Introduction

Zampoukas, 2011). In the European Union, the diversity of the phytoplankton community is used as a biological indicator of the ecological status of water bodies monitored in accordance with the Water Framework Directive (European Parliament Council, 2000). In addition, the World Health Organization (WHO) has established guideline values for drinking-water supplies and recreational waters which may contain toxic cyanobacterial populations (Chorus, 2005). Such monitoring is not only used to assess the evolution of the enviromental status of aquatic ecosystems but also to understand ecological processes, including the growth and succession of phytoplankton species.

The complexity of the processes involved in the development of algal blooms cannot totally be studied with costly mesocosm or microcosm approaches (Carpenter, 1996; Liu et al., 2007). Ecological models are able to represent different ecosystem processes that control algal blooms; they help better understand how they interact and seek their responses to different inputs. Ecological models enable researchers to simulate and analyze ecological processes in an ecosystem over long periods of time, from months to decades. Another major use of models is as predictive tools supporting inter-disciplinary ecosystem management (Carpenter et al., 1999). They can be used by managers to predict future states of the ecosystem. It helps preview the consequences of different management decisions, e.g. in response to changes in the climate, in the land use in the catchment, or for biodiversity protection.

Lake models which describe eutrophication processes were developed for environmental management already in the 1970s (Jørgensen, 2008). Models with a wide spectrum of complexity were applied for environmental management since the 1980s (Bormans and Condie, 1997; Gaillard, 1981; Jørgensen, 2010; Salençon, 1997). These models vary in their dimensionality and the number of processes they cover. The recent process-based models described in the literature are often highly complex and require large data sets (Bruce et al., 2006; Gal et al., 2009). Beyond a certain degree, adding processes was shown to reduce model predictive capabilities (McDonald and Urban, 2010; Mieleitner and Reichert, 2008). Lake and reservoir managers are in need of simple tools for understanding the drivers of harmful algal blooms and predicting them.

Many lakes and reservoirs in the Middle East are poorly documented, except for Lake Kinneret in Israel. Karaoun Reservoir is the largest freshwater body in Lebanon, with a

maximum capacity of 224×106 m3. There is a great interest in the water quality of this reservoir since its waters are anticipated to supply Beirut with drinking water (Greater Beirut Project). The few documents which report the occurrence of toxic cyanobacterial blooms in Karaoun Reservoir do not describe the physico-chemical factors that control these blooms (Atoui et al., 2013; Slim et al., 2013). Despite the interest of Karaoun Reservoir, its hydrodynamic and ecology was never modelled before.

Thesis objectives and methodology

The main objective of this Ph.D. work is to understand the dynamics of phytoplankton in Karaoun Reservoir. This main objective branches into three sub-objectives: 1) to establish the seasonal phytoplankton succession 2) to understand the cyanobacterial dynamics and 3) to identify the driving factors of the cyanobacterial blooms.

To achieve these objectives, the work plan was conducted in two directions:

1) to design and implement a physico-chemical and reinforced biological monitoring in Karaoun reservoir;

2) to implement a simple deterministic model for a better understanding of the drivers of cyanobacteria blooms.

More precisely, in the first period of the thesis, reliable existing data on Karaoun Reservoir was gathered and analysed in a review article on its geology, hydrology, uses and ecological status (Fadel et al., 2014). Field campaigns were then conducted on Karaoun Reservoir in 2012 and 2013. The analysis of the measurements allowed us to understand the succession and competition between phytoplankton groups and cyanobacterial species. They were then used to calibrate and verify the ability of a hydrodynamic-ecological model to simulate changes in water level, water temperature profiles and succession between toxic cyanobacterial species in Karaoun Reservoir.

Structure of the thesis

This dissertation consists of 5 chapters and 4 appendices. Chapter 1 starts by presenting the actual and future development of reservoirs and their ecosystems. It then discusses the development of toxic cyanobacterial blooms in freshwater bodies, their key controlling

General Introduction

factors and their ecological and health impacts. The chapter also gives an overview of the most commonly applied hydrodynamic-ecological models. Chapter 2 describes the study site and the methods used in this work. It starts by describing the geology, hydrology and current and anticipated uses of Karaoun Reservoir. It then presents the monitoring and laboratory analysis program that was designed to monitor the succession between phytoplankton groups and their physico-chemical conditions. In the end it presents the applied hydrodynamic-ecological model, DYRESM-CAEDYM2. In chapter 3, the trophic and biodiversity status of Karaoun Reservoir are assessed and the algal succession described, first before 2012 from literature data, second for the 2012-2013 study period, in more detail and from campaign results. Among this algal succession, chapter 4 focuses on the competition between the two main toxic cyanobacteria observed in the reservoir. It describes the presence of cylindrospermopsin toxin and discusses its dynamics. Chapter 5 presents the simple configuration of the model, based on few processes, used to describe the succession between toxic cyanobacterial species. The different limiting factors in this succession are analysed from simulation results. A general conclusion sums up the main results and draws suggestions and perspectives for further work.

The first period of the thesis was dedicated to field campaigns and modelling in France. Data of the ANR PROLIPHYC3 research project on Grangent reservoir (France) were processed to calibrate the DYRESM-CAEDYM model to simulate cyanobacteria dynamics in an early warning system. This research work is presented in an article entitled “A simplified model as a warning system of harmful algal blooms in lakes, application to Grangent reservoir, France”.

Supplementary elements concerning the field survey and laboratory analysis are presented in appendices A and C.

2

DYRESM - Dynamic Reservoir Simulation Model; CAEDYM – Computational Aquatic Ecosystem Dynamic Model. Centre for Water Research, University of Western Australia.

3 _{The Proliphyc system is an autonomous buoy that performs meteorological measurements and includes} profilers with underwater sensors

Chapter 1 Literature review

Chapter 1

Literature review

This chapter aims to introduce background knowledge used in this Ph.D work and to define key words of this research field. It gives an overview of the actual and future development of reservoirs, their physical functioning (thermal stratification and mixing). It describes cyanobacteria characteristics, the environmental factors controlling their development, their ecological and health impacts of cyanobacterial blooms. The last section presents the main hydrodynamic-ecological models published in the literature.

1.1

Reservoirs and their ecosystems

1.1.1

Actual and future development of reservoirs

Reservoirs are man-made water-bodies formed by constructing a dam across a flowing river to store water. Due to their ability to retain, store, and evenly provide water, lakes and reservoirs constitute essential components of the hydrological and biogeochemical water cycles, and influence many aspects of ecology, economy, and human welfare (Lehner and Doll, 2004).

The growing populations of cities and developing industry require greater quantities of water with each passing year. Over 45 000 large dams were constructed in over 140 countries by the end of the 20th century (WCD (World Comission on Dams), 2000). The worldwide development of reservoir storage in the 20th century was slow until the mid 1950s when large projects began to come on streams increasing the rate until 1980s (Figure 1). About 16.7 million reservoirs larger than 100 m2 with a combined storage capacity of approximately 8070 km3 exist worldwide (Lehner et al., 2011). Among them, 6862 dams and their associated reservoirs, with a total storage capacity of 6197 km3 are referenced in the Global Reservoir and Dam database (GRanD). A continuous increase in the construction of reservoirs on rivers is expected, in an attempt to address the increasing human needs for drinking, irrigation and hydropower (Seitzinger et al., 2010).

Chapter 1 Literature review 0 1000 2000 3000 4000 5000 6000 7000 1900 1920 1940 1960 1980 2000 Date W o rl d r es er v o ir c a p a ci ty ( k m 3 )

Figure 1 Development of worldwide reservoir storage since 1900 (White, 2010).

There are numerous reasons why water storage reservoirs are necessary. Here are some positive effects of reservoir construction:

1. Flood control and storage for further drinking and municipal water supply.

2. Production of hydroenergy which is ecologically considered as the cleanest form of energy. About 20% of the worldwide generation of electricity is attributable to hydroelectric schemes. This equates to about 7% of worldwide energy usage (White, 2010).

3. Irrigation of about 3.106 km2, which represents 20% of cultivated land worldwide, which produces 33% of the world food supply (White, 2010).

4. Creation of favourable conditions for tourism and recreational activities (fishing, boating, etc.).

Beside the positive effects, negative impacts result from reservoir construction:

1. Temperature of water, nutrients and oxygen distribution may change vertically as a consequence of reservoir formation. This may result in the development of exogenous living species, eutrophication and harmful algal blooms (Tahmiscioğlu et al., 2007).

Chapter 1 Literature review

2. Large amounts of plant life are submerged and decay anaerobically (in the absence of oxygen), generating greenhouse gases like methane (Chanudet et al., 2012).

3. Unexpected floods may occur and result in the displacement of local population and the damage of vegetation and natural structures in the riverbanks.

4. Usual passing ways of territorial animals are hindered since the dam works as a barrier. Also, the migratory pattern of river animals like salmon and trout are affected (Stott and Smith, 2001).

1.1.2

Differences between lakes and reservoirs

Reservoirs receive through rivers larger inputs of water, as well as soil and pollutant loads than lakes. On the other hand, reservoirs have the potential to flush the pollutants more rapidly than do lakes through water withdrawal (UNEP, 2000).

Unlike lakes, deep reservoirs are distinguished by the presence of a longitudinal gradient in physical, chemical and biological water quality characteristics from the upstream river end to the dam. Thus, reservoirs have three major zones: an upstream riverine zone, a downstream lake-like zone at the dam end, and a transitional zone separating these two zones (Figure 2, Table 1).

Chapter 1 Literature review

Table 1 Difference between the three major zones of deep reservoirs (UNEP, 2000).

Riverine zone Transitional zone Lacustrine zone

Narrow basin Board basin Board basin

High flow rates Reduced flow rates Low flow rates High suspended solids Reduced suspended solids Low suspended solids Low light and high nutrient

availability

Higher light and lower nutrient availability

Highest light and lowest nutrient availability Primary productivity limited

by light

High primary productivity Primary productivity limited by nutrients

1.1.3

Physical functioning of lakes and reservoirs

1.1.3.1 Development of thermal stratification

Lakes and reservoirs are affected by different meterological (radiation, precipitation and wind) and hydrological processes (inflows, outflows) (Figure 3). Thermal stratification is a major factor influencing the growth and succession of phytoplankton and overall water quality in lakes. It is a solar-radiation-driven process that separates the lake water column into three distinct vertical layers due to the change in water density with temperature (Lampert and Sommer, 2007). Radiation mainly enters the lake at short wavelengths (visible light); it is mostly absorbed near the surface and transformed into heat. Water density is maximum at 4 °C at low turbidity and salinity; it becomes lighter either by cooling below 4 °C or by warming above 4 °C. Lighter water buoys whereas denser water sinks. Wind produces turbulence and currents at the surface, that dampen rapidly with depth, creating a transition layer, the metalimnion, between the mixed warm surface waters, the epilimnion, and the colder quiescent deep waters, the hypolimnion. This yields the typical temperature profile of a stratified lake of long residence time with a strong decline in temperature in the metalimion, called thermocline (Figure 4), separating the epilimnion and the hypolimnion, both with rather homogeneous temperatures.

Chapter 1 Literature review

Figure 3 Different meteorological (radiation, evaporation, precipitation and wind) and hydrodynamic processes that affect reservoirs , adapted from (Hipsey et al., 2012).

Stratified lakes are separated into three distinct vertical zones (Figure 4):

- the epilimnion, the upper well mixed layer, is marked by little variation in temperatures over depth.

- the metalimnion, a transition layer, between the epilimnion and the hypolimnion, temperatures decrease rapidly with depth. This layer contains the thermocline, a thin but distinct layer in which temperature changes more rapidly with depth than it does in the layers above or below. The location of the thermocline varies with the depth of the outlet used for withdrawal in reservoirs (Casamitjana et al., 2003). - the hypolimnion, the bottom layer, below the metalimnion, shows little change in

Chapter 1 Literature review

Figure 4 Typical temperature profile in a stratified lake (Bade, 2005)

This type of stratification does not apply to all lakes and reservoirs. Some reservoirs may only have an epilimnion and a metalimnion (Bade, 2005). The location of the thermocline is not governed by solar radiation and wind only, but varies with the depth of the outlet used for withdrawal (Casamitjana et al., 2003) and can also be influenced by inflows.

1.1.3.2 Stratification and mixing patterns in lakes and reservoirs

Different patterns of seasonal stratification and mixing in lakes are observed throughout the world (Lewis Jr., 1983):

1. Amictic lakes never mix because they are permanently frozen. They exhibit inverse cold water stratification whereby water temperature increases with depth below the ice surface 0 °C. Such lakes are found in Arctic and Antarctic regions and at very high altitudes.

2. Meromictic lakes mix only partially; the deep water layers never intermix either because of high water density caused by dissolved substances or because the lake is protected from wind effects. An example of a meromictic lake is Lake Pavin (Bonhomme et al., 2011).

Chapter 1 Literature review

3. Holomictic lakes mix completely at some time in the year; their temperature and density is uniform from top to bottom at that time. These lakes are classified according to the frequency of mixing:

– Oligomictic lakes do not mix every year. Because such lakes are usually large and have a large heat storage capacity, whether or not they mix completely depends on the local meteorological conditions. An example is Lake Bourget in France (Vinçon-Leite et al., 2014).

– Monomictic lakes mix only once each year, either in summer or in winter: i. Cold monomictic lakes are found in Polar regions. They are covered by

ice throughout much of the year. They thaw, but rarely reach temperatures above 4 °C, and mix in summer.

ii. Warm monomictic lakes mix in winter. These lakes are widely distributed from temperate to tropical climatic regions. One example is Lake Constance, which on average freezes over about once every 33 years because of its large size.

– Dimictic lakes mix twice a year (usually in spring and autumn). During winter they are covered by ice. During summer they are thermally stratified, with temperature-derived density differences separating the warm surface waters (the epilimnion), from the colder bottom waters (the hypolimnion). This is the most common lake type at temperate latitudes.

– Polymictic lakes mix frequently. These are usually shallow lakes that do not develop seasonal thermal stratification. Their stratification can last a few days or weeks. They are found both under tropical and temperate latitudes. An example is Lake Créteil in France (Soulignac et al., 2014).

This classification is based on mixing regimes in lakes rather than reservoirs that are generally classified as polymictic.

1.1.3.3 Impact of stable stratification on water quality and biodiversity

Thermal stratification is a major factor influencing the growth and succession of phytoplankton and overall water quality in lakes. During stratification period, mixing is dramatically reduced in the hypolimnion, compared to the epilimnion and to a fully mixed

Chapter 1 Literature review

lake. Oxygen conditions, nutrient cycling and phytoplankton biomass are affected by this reduced mixing (Huisman et al., 2004; Straile et al., 2003): vertical mixing of oxygen from the lake surface is hindered, which can lead to hypoxia or even anoxia in the hypolimnion, depending on the duration of stratification and on water temperature (Jankowski et al., 2006; Wilhelm and Adrian, 2008).

Thermal stratification favours buoyant species like cyanobacteria over denser species like diatoms (Huisman et al., 2004). Moreover, thermal stratification isolates the epilimnion from the nutrient-richer bottom layers, resulting in phosphorus depletion in the epilimnion. This increases the occurrence of motile algal species such as the dinoflagellate Ceratium hirundinella (Anneville et al., 2002) or the cyanobacterium Planktothrix rubescens (Micheletti et al., 1998). Due to their large body size, these algae are better protected from zooplankton grazing and may build up high standing stocks by the end of the stratification period.

Mixing is an essential process that prevents the build up of high phosphate gradients at the sediment-water interface in the hypolimnion (Marsden, 1989). However, a stratified water column hampers the phosphates released from the sediment in the hypolimnion and prevents their release to the euphotic zone where photosynthesis takes place (DeStasio et al., 1996). The breakdown of thermal stratification after the autumn overturn allows nutrient release into the euphotic zone and may induce phytoplankton blooms.

1.2

Cyanobacteria in freshwater bodies

Cyanobacteria, also known as Cyanophyta, range from unicellular to unspecialized colonial aggregations and form the most widely distributed group of phytoplankton (Reynolds, 2006a). They can be benthic or planktonic, occasionally forming blooms in eutrophic lakes, and are an important component of the picoplankton in both marine and freshwater systems. Their pigmentation includes chlorophyll a, blue and red phycobilins (phycoerythrin, phycocyanin, allophycocyanin, and phycoerythrocyanin) and carotenoids (Barsanti and Gualtieri, 2006). Cyanobacteria include four main orders, of which three have planktonic representatives (Reynolds, 2006a):

Chapter 1 Literature review

1. Chroococcales: Unicellular or colonial cyanobacteria but never filamentous. Most

planktonic genera form mucilaginous colonies (assemblies of cells linked by a viscous exsudate called mucilage), and these are mainly encountered in fresh water. Picophytoplanktonic forms are abundant in the oceans. This order includes Aphanocapsa, Aphanothece, Chroococcus, Cyanodictyon, Gomphosphaeria, Merismopedia, Microcystis, Snowella, Synechococcus, Synechocystis, and Woronichinia.

2. Oscillatoriales: Uniseriate–filamentous cyanobacteria whose cells all undergo

division in the same plane. Marine and freshwater genera. This order includes Arthrospira, Limnothrix, Lyngbya, Planktothrix, Pseudanabaena, Spirulina, Trichodesmium and Tychonema.

3. Nostocales: Unbranched–filamentous cyanobacteria whose cells all undergo division

in the same plane and certain of which may be facultatively differentiated into heterocysts. Heterocysts are vegetative cells that have been drastically altered (loss of photosystem II, development of a thick, glycolipid cell wall) to provide the necessary anoxic environment for the process of nitrogen fixation. This order includes Anabaena, Anabaenopsis, Aphanizomenon, Cylindrospermopsis, Gloeotrichia and Nodularia.

1.2.1

Ecological and health impacts of toxic cyanobacterial blooms

When they bloom, cyanobacteria have adverse impacts on aquatic ecosystems and human health, with wide-ranging economic and ecological consequences. According to many authors, climate warming is expected to favour the dominance of cyanobacteria on other phytoplankton communities (Paerl and Paul, 2012). An improved understanding of the interactions amongst both the environmental drivers and cyanobacterial physiology is necessary to develop strategies to reduce the risk of more frequent blooms (Brookes and Carey, 2011).

1.2.1.1 Impacts on ecosystems

The proliferation of cyanobacteria can have numerous consequences. In addition to economic costs for water treatment and losses in tourism, property values, and business (Dodds et al.,

Chapter 1 Literature review

rise pH. The formation of cyanobacterial blooms in the epilimnion in high light intensities, increases turbidity and decreases light availability (Nõges and Solovjova, 2005) for other primary producers. The increase in pH during intense cyanobacterial blooms may be harmful to certain species of fish (Kann and Smith, 1999). Summer cyanobacterial blooms that lasted for weeks in Chesapeake Bay (USA) increased pH to 10.5; this elevated pH promoted desorption of sedimentary inorganic phosphorus (Gao et al., 2012).

Figure 5 Common effects of cyanobacterial blooms.

Oxygen depletion can be due to the high respiration of cyanobacteria at night or in dim light during the day or to the aerobic degradation of cyanobacteria during bloom decay. This causes severe biological impacts like fish kills (Hallegraeff, 1993). Between 1951 and 2006, cyanobacterial blooms and hypoxia killed more than 383 million fish in Texas (Thronson and Quigg, 2008). Moreover, oxygen depletion has other effects including the release of nutrients and heavy metals from the sediment. Nutrients and toxic heavy metals are highly concentrated in the sediment. When the dissolved oxygen concentration lowers, nutrients and metals

Cyanobacterial

blooms

Reduce water

transparency

Produce toxins and

reduce biodiversity

Increase pH

Decrease oxygen / Hypoxia

after bloom decay

Produce bad

odors

Death of living organisms due to

toxins and/or hypoxia

Chapter 1 Literature review

including toxic heavy metals are released due to the chemical reduction of oxides. This can lead to further cyanobacterial blooms and to the contamination of the water column.

1.2.1.2 Toxin production

Many cyanobacterial species produce cyanotoxins that affect animals and humans. Cyanotoxins are a diverse group of natural toxins, both from the chemical and the toxicological points of view. They are classified according to how they affect the human body (Table 2). Here we present the two most widespread cyanotoxins, microcystin and cylindrospermopsin.

Chapter 1 Literature review

Table 2 Principle groups of cyanobacterial toxins and their sources (Araoz et al., 2010; Chorus, 2005; Codd et al., 2005)

Toxin Structure Activity Toxigenic genera

Microcystins Cyclic heptapeptides Protein phosphatase- inhibition, membrane integrity and

conductance disruption, tumour promoters

Microcystis, Anabaena, Nostoc, Anabaenopsis, Planktothrix, Oscillatoria, Hapalosiphon

Nodularins Cyclic pentapeptides Protein phosphatase-inhibition, membrane integrity and

conductance disruption, tumour promoters, carcinogenic

Nodularia, Theonella

(sponge-containing cyanobacterial symbionts)

Hepatotoxins

Cylindrospermopsins Guanidine alkaloids necrotic injury to liver (also to kidneys, spleen, lungs, intestine), protein synthesis inhibitor,

genotoxic

Cylindrospermopsis, Aphanizomenon, Umezakia, Anabaenac, Raphidiopsisd

Anatoxin-a (including homoanatoxin-a)

Alkaloids postsynaptic, depolarising neuromuscular blockers

Anabaena, Oscillatoria, Phormidium, Aphanizomenon, Rhaphidiopsise

Neurotoxins

Saxitoxins Carbamate alkaloids sodium channel-blockers Aphanizomenon, Anabaena, Lyngbya, Cylindrospermopsisf s ,

Planktothrix

Lyngbyatoxin-a Alkaloids Inflammatory agents, protein kinase C activators

Lyngbya, Schizothrix, Oscillatoria

Dermatotoxins and cytotoxins

Aplysiatoxins Alkaloids Inflammatory agents, protein kinase C activators

Lyngbya, Schizothrix, Oscillatoria

Endotoxins Lipopolysaccharides Lipopolysaccharides inflammatory agents, gastrointestinal irritants

Chapter 1 Literature review

1.2.1.2.1 Microcystins

Microcystins are of the most frequent cyanobacterial toxins that can be found in freshwater bodies. These cyclic non ribosomal peptides are produced by multiple genera of cyanobacteria.

Figure 6 Structure of microcystin-LR (Leucine-Arginine) (An and Carmichael, 1994)

In lakes, microcystins are mostly found inside cyanobacteria cells. However, they can be released out of the cells during cell lysis in the decline of blooms and breakdown of biomass (Codd et al., 1999). Microcystins are very stable compounds, they can withstand many hours of boiling at temperature 300°C at neutral pH (Harada et al., 1996), their photodegradation by sunlight is negligible (Tsuji et al., 1995) and they may persist for many years if stored dry at room temperature (Svrcek and Smith, 2004).

Consumption of microcystin-contaminated water can cause liver damage and even human death. In 1996, 50 people died in Caruaru, Brazil, due to the use of hemodialysis water contaminated by microcystin (Pouria et al., 1998). Microcystin has the potential to promote cancer in humans even at low concentrations (De Figueiredo et al., 2004). For that, the World Health Organization (WHO) has established a provisional guideline value of 1 µg L-1 microcystin-LR in drinking water (WHO, 1998). Microcystin-LR (MCLR) has several analogs; the most documented are microcystins LA (MCLA), LF (MCLF), LR (MCLR), -LW (MC-LW) and -RR (MCRR).

Chapter 1 Literature review

1.2.1.2.2 Cylindrospermopsin

Cylindrospermopsin (CYN) is a water soluble alkaloid hepatotoxin produced by several toxic cyanobacteria (Table 3, Figure 7). It causes damage to kidney, lungs and heart. It was also reported as protein synthesis inhibitor, genotoxic (Humpage et al., 2000) and carcinogenic (Falconer and Humpage, 2006). It is believed to be responsible for the severe hepatoenteritis that affected 148 people in 1979 on Palm Island, Queensland, Australia (Griffiths and Saker, 2003).

Figure 7 Structure of cylindrospermopsin

Table 3 Geographical distribution of freshwater cylindrospermopsin-secreting cyanobacteria

Cyanobacteria species Country Reference

Anabaena bergii Australia (Schembri et al., 2001)

Anabaena lapponica Finland (Spoof et al., 2006)

Anabaena planctonica France (Brient et al., 2009)

Cylindrospermopsis raciborskii USA (Chapman and Schelske, 1997)

Aphanizomenon flos-aquae Germany (Preußel et al., 2009)

Aphanizomenon ovalisporum Israel (Banker et al., 1997)

Lyngbia wollei Australia (Seifert et al., 2007)

Raphidiopsis curvata China (Li et al., 2001)

Chapter 1 Literature review

A larger fraction of cylindrospermopsin (up to 96%) is found as extracellular form (Bormans et al., 2014). It persists in many water bodies because of its chemical stability and slow degradation (Wörmer et al., 2008). Recently, it was found in dangerous concentrations in many freshwater bodies through out the world. In Germany, concentrations up to 12.1 µg L-1 were reported in lakes (Rücker et al., 2007); in Spain, they reached 9.4 µg L-1 (Quesada et al., 2006), and 18.4 µg L-1 in Italy (Bogialli et al., 2006).

1.2.2

Main functional traits and key controlling factors of

cyanobacterial blooms

The distribution and persistence of species composition that are often observed in phytoplankton communities in lakes and reservoirs result from a complex interplay between physical and chemical properties of the aquatic environment (nutrient amount, nutrient ratios, turnover speed, temperature, water density, light regime) on the one hand (Kangro et al., 2005; Smith and Benne, 1999)and the response of the individual species on the other hand (growth rate, internal utilization, transport and storage of nutrient, optimum temperature and irradiance, species mobility) as well as the grazing rate (Dignum et al., 2005); (Kõiv and Kangro, 2005).

1.2.2.1 Temperature

Water temperature affects a number of physical, chemical, and biological processes in natural aquatic systems (Ibelings et al., 2011). It affects the metabolism, growth, reproduction, and survival of living organisms, and therefore the interactions among species (Kingsolver, 2009). An increase in water temperature increases the solubility of chemical compounds (nutrients) and decreases the solubility of oxygen participating in its depletion.

Beside nutrient and light, temperature can limit cyanobacterial growth (Davison, 1991; Zheng et al., 2008) as it directly controls the photosynthetic capacity, specific respiration rate and the replication rates of phytoplankton communities (Robarts and Zohary, 1987; Salmaso et al., 2012). Water temperature is an important controlling factor for both bloom development and seasonal succession. Blooms can persist in waters with temperatures between 15 and 30°C, with maximum growth rates occurring at temperatures in excess of 25°C. Laboratory

Chapter 1 Literature review

nutrient and light-concentrated conditions; 0.083/day at 15 °C, 0.42/day at 20 °C and 0.81/day at 30 °C (Chu et al., 2007).

Water temperature affects the buoyancy of Microcystis sp.; non-buoyant Microcystis colonies taken from a small eutrophic pond near Bristol regained buoyancy in the dark rapidly at 20 °C but only slowly at 12 °C and below (Thomas and Walsby, 1986). Indeed, low temperature conditions (between 8 and 12 °C) depress carbohydrate metabolism that affects the buoyancy regulation of Microcystis (Oliver and Gnaf, 2002).

1.2.2.2 Light

Light is an essential resource for all photoautotrophic organisms and its availability has a large influence on phytoplankton composition and diversity (Floder et al., 2002). The intensity of daylight needed to optimise growth depends on the cyanobacterial species. Planktothrix dominates under low irradiance conditions while Anabaena, Aphanizomenon and Microcystis dominate during higher irradiance conditions (Havens et al., 1998).

Cyanobacteria can tolerate low irradiance by enhancing cell-specific photosynthetic potential. It increases the cell-specific light-harvesting capacity, by increasing the number of light harvesting chlorophyll-protein complexes (LHCs) in individual cells (Reynolds, 2006d). Although light is essential for photosynthetic organisms, an extended exposure to high light intensities is lethal for many species (Wu et al., 2011). Too high intensities can degrade the photosynthetic apparatus of cyanobacteria (photo-inhibition) and even cause their degeneration at very high intensities. However, cyanobacteria have a competitive advantage over other algae. They can escape photo-inhibition and sink down to low irradiance depth (Paerl et al., 1985).

1.2.2.3 Nutrients

Nutrients are essential for the growth of phytoplankton. Nitrogen exists as dissolved, particulate and gaseous forms. Nitrate and nitrite, and ammonia nitrogen are forms that are biologically available and readily exchanged within and between the water column and the sediment. However, some heterocystous cyanobacteria species are able to fix atmospheric nitrogen (Hadas et al., 1999; Yamamoto, 2009). Nitrogen-fixing cyanobacteria (e.g.,

Chapter 1 Literature review

Aphanizomenon sp., Anabaena sp. and Cylindrospermopsis sp.) benefit from a competitive advantage when nitrogen sources are strongly depleted in the water column (Wood et al., 2010).

Cyanobacteria can assimilate N significantly without comparable P uptake during the blooming season (Ahn et al., 2002). The availability of N in summer was found as the key growth-limiting factor for the initiation and maintenance of toxic non-nitrogen-fixers, such as Microcystis aeruginosa in Copco and Iron Gate Reservoirs in the Klamath River (Moisander et al., 2009).

In many freshwater systems, phosphorus (P) is a limiting nutrient (Schindler et al., 2008). It plays a predominant role in cells. It is an energy and information carrier (Parkinson and Kofoid, 1992) and a basic element in cell wall and DNA (Jürgens et al., 1983). Total phosphorus represents the sum of all forms of phosphorus: dissolved and particulate organic phosphorus from algae and other organisms and from mineral particles. Orthophosphate is the only directly available phosphorus source for phytoplankton (Palenik et al., 2003). Phosphorus is often the limiting factor of phytoplankton growth because of its low concentrations within the water column. But phytoplankton can stay productive even if the amount of phosphorus is reduced to 0.3 µg L−1 (Hudson et al., 2000). Cyanobacteria can outcompete other phytoplankton organisms under conditions of nutrient limitation, due to their high nutrient affinity and storage capacity (Carey et al., 2012). Low nutrient concentration in the epilimnion can force the phytoplankton community to change to species that can tolerate high light and low nutrient conditions (Sommer et al., 1986).

The amounts, proportions and chemical composition of N and P sources influence the composition, magnitude and duration of blooms (Bulgakov and Levich, 1999). Cyanobacteria dominate at low N:P ratios and temperatures higher than 15 °C (Havens et al., 2003; Lilover and Stips, 2008; Ndong et al., 2014).

1.2.2.4 Vertical migration

Vertical migration allows some phytoplankton species to move down and upwards through water layers. It is achieved using flagellates in motile organisms and intracellular gas vesicles that regulate buoyancy. Vertical migration permits cyanobacteria to access to nutrient supplies

Chapter 1 Literature review

that are available in the deep layers (Wagner and Adrian, 2009). The regulation of the migration velocity is achieved through a balance between buoyancy through the gas vesicles and sinking through the carbohydrate stores acting as ballast (Walsby, 1994). Buoyancy increases the supply of light and supports a higher rate of photosynthesis as buoyant cells move upwards to the illuminated surface layers (Humphries and Lyne, 1988; Walsby et al., 1989). The obvious benefit of buoyancy is demonstrated after deep mixing episodes that move phytoplankton below the thermocline. In such case and once stratification is established, buoyant cells can move to the surface while non-buoyant cell continue to settle (Humphries and Lyne, 1988).

Cyanobacteria migration velocity may reach 40-60 µm/s as in the case of Anabaena and Aphanizomenon and 100-300 µm/s for large Microcystis colonies (Oliver, 1994; Reynolds et al., 1987). This later velocity range is comparable to that of the larger dinoflagellates, such as freshwater Ceratium and Peridinium, 200–500 µm/s (Reynolds, 2006b).

1.2.2.5 Grazing

Phytoplankton is generally vulnerable to severe physical biomass losses by many herbivores that inhabit water bodies as zooplankton, zebra mussels and planktivorous fish. Grazing can cause a loss in cyanobacteria but they are generally less grazed than other phytoplankton groups (Sarnelle, 2007; Yang et al., 2008; Zhang et al., 2009). Large-sized Daphnia can consume small-sized cyanobacteria colonies with a diameter smaller than 50 µm but they have difficulties in ingesting large colonies. Some colony-forming cyanobacteria like Microcystis aeruginosa can produce mucilage that has an inhibitory effect on swallowing and consequently on the ingestion rate of these cyanobacteria by Daphnia sp. (Rohrlack et al., 1999). The grazing rate of cyanobacteria depends on the duration of the presence of zooplankton and the presence of fish that feed on zooplankton (Sarnelle, 2007).

1.2.2.6 Wind mixing and flushing

Storms count among the main factors that can cause major dispersion and loss in phytoplankton groups. Trimbee and Harris (1983) observed that physical changes in a lake appeared after a delay of one day in response to a strong wind event and that the biological change occured after a delay of 2 to 3 days after the wind. Intense wind speed can destroy the